This study investigated the construct validity of the revised Special Education-Mental Handicaps Georgia Teacher Certification Test (MH-TCT) using hierarchical confirmatory factor analysis and LISREL VI. The primary objective was to determine whether first-order and second-order factors correspond to item/objective/test relationships defined by judgments/consensual content validation studies. Data from four administrations of the MH-TCT were used in the analyses. Subjects included 287 candidates for certification in the mental handicaps subarea of special education. Since the sample size was small, findings were tentative. LISREL was used to examine the factor structure of the entire test as well as to identify test items that were not psychometrically sound. Although six subdimensions were identified that correspond to the six test objectives, first-order factors corresponding to these were so highly correlated that a hierarchical model was rejected in favor of a unidimensional model. A unidimensional model was invariant across samples and time. Results suggest that the LISREL confirmatory maximum likelihood factor model provided a powerful tool for item analysis and test validation. A major drawback was the amount of computer run time needed to do maximum likelihood factor analysis. Four data tables and one flowchart are provided. (Author/TJH)